TrigunAI01 · Parents
Try it with your child
अपने बच्चे के साथ आज़माइएSit with your child for one short session. Then tell us how your child’s studies go at home in our 10–15 minute survey.
TrigunAI · a research-based product company
We study the real problem before we build.
We work like a small frontier lab. Research is the backbone, and it decides what we build. No product ships until the research shows the real problem.
Join the Acharya beta
Parents: take the survey
Free research beta of our adaptive tutor · माता-पिता के लिए 10–15 मिनट का सर्वे · nothing to buy
Pratima
Foundational research on the human body and how it moves. Groundwork, not a product.
Acharya
Education research. Verified tests that find a learner’s weak points, and an adaptive tutor in free research beta.
Open beta · Acharya adaptive tutor · बीटा अब खुला है
Help us find out whether adaptive learning helps.
This beta is research, not a launch. The tutor is not a finished product and costs nothing. We are asking for your use and your honest feedback.
Does finding a child’s exact gap first, then teaching it in tiny steps, help more than an ordinary lesson? And does it still hold a week later?
Acharya walks down a map of maths concepts until it finds the one a child is ready to learn next, teaches that in small steps the child does with their own hands, and checks a week later that it stayed. Whether that actually helps is the open question. What you do in it, and what you tell us, is how we will answer it.
- Start with a test: take a verified, exam-pattern test on evaluate.trigunai.com. It maps your strengths and weak topics, and the map sharpens the more tests you take.
- Then the tutor sets up from it: that weak-point map is what we are wiring into the tutor during the beta, so it starts where you actually are.
- Ready to try: maths for classes 2–3, in English, Hindi or Hinglish, with voice, on any phone.
- Still being built: more concepts, more classes, other subjects. Expect rough edges.
- What we will publish: only what our own one-week retests measure, including if it does not help.
Four kinds of people we want in the beta
02 · Students
Find your weak topics first
पहले अपनी कमज़ोरी पहचानिएTake verified, exam-pattern tests for JEE, NEET, boards, UPSC, SSC, banking and Bihar exams. Acharya maps what you know and what you don’t. Then try the tutor and tell us where a step lost you.
03 · Teachers & tutors
Check its diagnosis against yours
क्या इसकी पहचानी कमी आपकी पहचान से मिलती है?Use it with one or two students. Does the gap Acharya finds match the one you see in class? You know your students; that is the feedback we need most.
04 · Coaching institutes & schools
Run a small pilot group
एक छोटे समूह के साथ पायलटTest one batch on verified papers, then try the tutor with them, free. We share back what the tests and the one-week retests show: what held and what did not.
Research
Two lines of inquiry
Each keeps what is established apart from what we are still testing.
- Established
- Hypothesis, being tested
Pratima
Led by Avinash, co-founder and Chief Research Officer
Long-term foundational research
How a believable human body is built, and how it moves and behaves in real environments. The body is not a 3D mesh. It is a mathematical model of a human: its joints, proportions, mass and strength.
Build the human once, carefully, and any character can inherit it: in XR first, later a real humanoid. A physical body is designed outward from the human model, never backward from whichever robot happens to be available.
Rules the work follows
- No number is chosen. Every value is sourced, or derived by a stated rule.
- A hardware limit is never the reason the human model takes a value.
- Human data is for comparison only, with an ethics review before any work with people.
Physical AI groundwork
Avinash runs the experiments behind our Physical AI and robotics line: a record of what actuators and sensors can do, a simulation adapter and a lab specification, so that the gap to hardware can be measured, not guessed. No humanoid has been built and no robot platform has been chosen.
Acharya
Led by Deepak, co-founder, CEO and CTO
User research · tutor in free open beta
Free YouTube and free AI already reach most households. Where does a learning gap remain that access to content does not close, and for whom?
The honest answer may be “for nobody we can serve”. The study is built so that it can say so.
Lead hypothesis
- Parents cannot see, judge or act on what their child is learning. A hypothesis to test, not a fact. It is measured against competing explanations (teaching quality, focus, the home, teacher shortage, structure, language, pace), so the data can prove it wrong.
What we are doing
- Field study, Patna. A parent survey about study at home, online and in person. Nothing is sold in the room.
- Open beta of the adaptive tutor. Maths, classes 2–3, set up from a verified test’s weak-point map. Does finding the gap first help, and does it last a week? Join it.
- Desk research. A first pass gathered 327 pieces of evidence from 186 sources. None of it is verified yet; a row is checked against its source before a decision quotes it.
Learning science
Find the exact gap first, teach in tiny steps the child does rather than watches, and count only what is still known a week later.
- Small steps, with practice after each one.
- Doing over watching; guided practice over discovery for beginners.
- Retrieval practice and spacing.
- Whether tiny steps beat ordinary lessons on what a child still knows after 7 days.
Method
Research is the backbone
Four rules hold for every line and every decision.
- 01
Claims are checked before we act
A claim may cite a source only when the source resolves, the exact words are found in it, and they support the claim. Gaps in evidence are counted, never guessed.
- 02
Hypotheses are labelled
What is established and what is being tested are marked apart, by shape as well as colour. A hypothesis never quietly becomes a fact.
- 03
Progress is what lasts a week
Same-day scores can reverse a week later. Learning counts only when it survives a retest after 7 days.
- 04
No product before the evidence
A study ends with what is broken, the evidence for it and what to test next. Only then do we decide whether to build, and what.
What you will not see from us: “learning styles”, same-day scores as progress, numbers we have not measured.
For parents in Patna · पटना के माता-पिता
घर पर पढ़ाई कैसे चलती है?
How do children’s studies go at home?
हम समझना चाहते हैं कि घर पर बच्चों की पढ़ाई कैसे चलती है — क्या ठीक चलता है और कहाँ मुश्किल आती है। हम कुछ भी बेच नहीं रहे हैं। इसमें लगभग 10–15 मिनट लगते हैं।
We want to understand how children’s studies go at home: what goes fine and where it gets difficult. We are not selling anything. It takes about 10–15 minutes.
What we collect, and why · हम क्या लेते हैं, और क्यों
- आपका नाम नहीं पूछा जाएगा। आपके जवाब निजी रहेंगे।We do not ask your name. Your answers stay private.
- आप कोई भी सवाल छोड़ सकते हैं या कभी भी रुक सकते हैं।You can skip any question or stop at any time.
- जवाब सिर्फ़ इस रिसर्च के लिए हैं — आपको कुछ बेचने के लिए नहीं।Answers are used only for this research, never to sell you anything, and are not shared outside TrigunAI.
- फ़ोन नंबर तभी, जब आप ख़ुद आगे बात करना चाहें — और वह जवाबों से अलग रखा जाता है।A phone number only if you choose to talk further, kept apart from your answers.
In line with India’s Digital Personal Data Protection Act, we ask your consent first and collect only what the research needs. See our privacy policy.
Online survey · open
सर्वे ऑनलाइन भरें — अपने फ़ोन पर।
Fill it in online, on your phone, in Hindi or English. About 10–15 minutes.
सर्वे भरें · Take the surveyAlso open: try our adaptive tutor with your child, free, as part of the research.
Questions about the study? deepak@trigunai.com
Team
Who leads the research
Avinash
CO-FOUNDER · CHIEF RESEARCH OFFICER
Leads Pratima: Physical AI and the human model.
avinash@trigunai.com